"Some students asked whether there was a fundamental contradiction between telling students to use online resources, but not to discuss the test with each other."
No.
No.
If the test is open book, open Internet, that doesn't mean you can ask your friends on Facebook for help, just because Facebook is on the Internet.
I think that the class was likely poorly designed, but that doesn't mean that the students are all in the clear.
That process is completely cut short if you just copy someone else's answer.
The implied intent of the rule is that READ-ONLY NON-INTERACTIVE Internet use was permitted. To today's students, the Internet is a READ-WRITE INTERACTIVE tool, unless explicitly stated otherwise.
Can you quote the rule because in the article it appears that the intent of the rule was that above all you couldn't interact with other students [on the course] to complete the work but as long as you kept to that rule then general internet resources were fine to use?
Personally, I think some of the most valuable learning happens when discussing course content with my peers. Sometimes, I take the role of the learner, other times I'm forced to test my own understanding when explaining things to others. It's a win-win. Because I engage with others, it's likely that many of the same ideas I put down on paper would be similar to other people's work. But I don't consider that cheating. However, I would never take an answer that somebody else had written, change the words around, and submit it. Not being directly involved in this case at Harvard, it's hard to tell exactly what the students did.